ctp Is There Really a Teacher Shortage? Center for the Study of Teaching and Policy

Transcription

1 Is There Really a Teacher Shortage? A Research Report co-sponsored by Center for the Study of Teaching and Policy and The Consortium for Policy Research in Education by Richard M. Ingersoll University of Pennsylvania September 2003 (Document R-03-4) ctp Center for the Study of Teaching and Policy U N I V E R S I T Y O F W A S H I N G T O N

2 ctp Center for the Study of Teaching and Policy A National Research Consortium U N I V E R S I T Y O F W A S H I N G T O N (lead institution) S T A N F O R D U N I V E R S I T Y T E A C H E R S C O L L E G E / C O L U M B I A U N I V E R S I T Y U N I V E R S I T Y O F M I C H I G A N U N I V E R S I T Y O F P E N N S Y L V A N I A Other active participants in CTP s research and dissemination program include researchers affiliated with Indiana University, Michigan State University, Pennsylvania State University, the University of California at Santa Barbara, the University of North Carolina, and Education Matters, Inc. CTP studies the way policies and conditions in schools, districts, states, and teacher education institutions shape the quality of teaching and learning in the nation s elementary and secondary schools. The Center pays particular attention to the ways these policies and conditions interact with each other to influence the teaching profession and its practice. The Center s program of research is carried out in collaboration with various other research organizations, among them other OERI-funded research centers, including the Consortium for Policy Research in Education (CPRE), the Center for Research on Education, Diversity, and Excellence (CREDE), and the Center on English Learning & Achievement (CELA). The Center is affiliated with a variety of professional and advocacy organizations that represent teachers, teacher educators, state and local policymakers, disciplinary groups, and educational reform interests. This research was supported by a grant from the Center for the Study of Teaching and Policy (CTP) under the Educational Research and Development Centers Program, PR/Award R308B970003, as administered by the National Institute on Educational Governance, Finance, Policymaking and Management, Office and Educational Research and Improvement, U.S. Department of Education. However, the contents do not necessarily represent the positions or policies of either national institute, OERI, or the U.S. Department of Education, or the endorsement of the federal government. This report draws from and builds on my earlier CTP report: Teacher Turnover, Teacher Shortages and the Organization of Schools, published in 2001 and also my earlier article: Teacher Turnover and Teacher Shortages: an Organizational Analysis, published in the fall 2001 issue of the American Educational Research Journal. Thanks are due to Emerson Elliott for helpful comments of this paper. Opinions in this report reflect those of the author and do not necessarily reflect those of the granting agencies.

3 TABLE OF CONTENTS Abstract...3 Introduction...5 The Data...6 Demand for Teachers Has Risen...6 How Adequate is the Supply of Teachers?...7 The Importance of Teacher Turnover for School Staffing Problems...9 The Revolving Door...11 The Importance of Teacher Turnover for Organizations...11 Teacher and School Differences in Turnover...13 The Sources of Teacher Turnover...15 Implications for Policy...17 Conclusion: What is to Be Done?...19 Endnotes...22 References...24 Appendix...28

4

5 ABSTRACT Contemporary educational thought holds that one of the pivotal causes of inadequate school performance is the inability of schools to adequately staff classrooms with qualified teachers. It is widely believed that schools are plagued by shortages of teachers, primarily due to recent increases in teacher retirements and student enrollments. This report summarizes a series of analyses that have investigated the possibility that there are other factors tied to the organizational characteristics and conditions of schools that are behind school staffing problems. The data utilized in this investigation are from the Schools and Staffing Survey and its supplement, the Teacher Followup Survey conducted by the National Center for Education Statistics. These data indicate that school staffing problems are not primarily due to teacher shortages, in the sense of an insufficient supply of qualified teachers. Rather, the data indicate that school staffing problems are primarily due to a revolving door where large numbers of qualified teachers depart their jobs for reasons other than retirement. The data show that the amount of turnover accounted for by retirement is relatively minor when compared to that associated with other factors, such as teacher job dissatisfaction and teachers pursuing other jobs. This report concludes that teacher recruitment programs traditionally dominant in the policy realm will not solve the staffing problems of such schools if they do not also address the organizational sources of low teacher retention. 3

6

7 INTRODUCTION Few educational problems have received more attention in recent years than the failure to ensure that elementary and secondary classrooms are all staffed with qualified teachers. Severe teacher shortages, it is widely believed, are confronting our elementary and secondary schools. We have been warned repeatedly that the nation will need to hire at least two million teachers over the next ten years (e.g., National Commission on Teaching, 1997, p ), and our teacher training institutions are simply not producing sufficient numbers of teachers to meet the demand. At the root of this school staffing crisis, according to the conventional wisdom, are two converging macro demographic trends increasing student enrollments and increasing teacher turnover due to a graying teaching force. The resulting shortfalls of teachers, the argument continues, force many school systems to resort to lowering standards to fill teaching openings, inevitably resulting in high levels of underqualified teachers and lower school performance. The prevailing policy response to these school staffing problems has been to attempt to increase the supply of teachers. In recent years a wide range of initiatives have been implemented to recruit new candidates into teaching. Among these are career-change programs, such as troops-to-teachers, designed to entice professionals into mid-career switches to teaching and Peace Corps-like programs, such as Teach for America, designed to lure the best and brightest into understaffed schools. Some school districts have resorted to recruiting teaching candidates from overseas. Many states have instituted alternative certification programs, whereby college graduates can postpone formal education training and begin teaching immediately. Financial incentives, such as signing bonuses, student loan forgiveness, housing assistance, and tuition reimbursement have all been instituted to aid teacher recruitment (Hirsch, Koppich & Knapp 2001; Feistritzer, 1997; Kopp, 1992). The No Child Left Behind Act passed in winter 2002 provides extensive federal funding for such initiatives. Teacher shortages and subsequent teacher recruitment initiatives are not new to the K-12 education system. In the early and mid 1980s a series of highly publicized reports trumpeted an almost identical series of diagnoses and prescriptions (see, e.g., National Commission on Excellence in Education, 1983; Darling-Hammond, 1984; National Academy of Sciences, 1987; for reviews of this issue, see Boe & Gilford, 1992). Indeed, teacher shortages have been a cyclic threat for decades (Weaver, 1983). Concern over teacher shortages in turn has spurred interest in empirical research on these issues, but until the past decade such efforts were limited by a lack of data. It was partly in order to address these data shortcomings that the U.S. Department of Education s National Center for Education Statistics conceived the Schools and Staffing Survey (SASS) and its supplement, the Teacher Followup Survey (TFS), beginning in the late 1980s (Haggstrom et al., 1988). This is now the largest and most comprehensive data source available on the staffing, occupational, and organizational aspects of schools. Over the past decade I have undertaken a series of research projects using SASS/ TFS to examine a range of issues concerned with teacher supply, demand, and quality (e.g., Ingersoll 1995, 1999, 2001a, 2003b). In this report I will summarize what these data tell us about the realities of school staffing problems and teacher shortages. The theoretical perspective I adopt in my research is drawn from organizational theory and the sociology of organizations, occupations, and work. My operating premise is that in order to fully understand the causes and consequences of these social problems it is necessary to examine them from the perspective of the organizations the schools and districts where these processes happen and within which teachers work. Employee supply, demand, and turnover are central issues in organizational theory and research. However, there have been few efforts to apply this theoretical 5

8 perspective to understanding school staffing problems and policy. As I will show, by bringing the organization back in, these school staffing problems are reframed from macro-level issues, involving inexorable societal demographic trends, to organizational issues, involving manipulable and policy-amenable aspects of particular schools. A close look at the data from this perspective, I argue, shows that the conventional wisdom concerning teacher shortages is largely a case of a wrong diagnosis and a wrong prescription. THE DATA As mentioned, the primary data source for this research is the nationally representative Schools and Staffing Survey (SASS) and its supplement, the Teacher Followup Survey (TFS), both conducted by the National Center for Education Statistics of the U.S. Department of Education. To date, four independent cycles of SASS have been completed: ; ; ; SASS is an unusually large survey. Each cycle of SASS administers survey questionnaires to a random sample of about 53,000 teachers, 12,000 principals, and 4,500 districts, representing all types of teachers, schools, districts and all 50 states. In addition, one year later, the same schools are contacted again, and all those in the original teacher sample who had moved from or left their teaching jobs are given a follow-up second questionnaire to obtain information on their departures. This latter group, along with a representative sample of those who stayed in their schools, comprise the Teacher Followup Survey. The TFS sample contains about 7,000 teachers. Unlike most previous data sources on teacher turnover, the TFS is large, comprehensive, nationally representative, and includes the reasons teachers themselves give for their departures and a wide range of information on the characteristics and conditions of elementary and secondary schools. It is also unusual in that it does not solely focus on a particular subset of total separations, but includes all turnover: voluntary, involuntary, transfers, quits, retirements, etc. In this report, I present data from all four cycles of SASS and TFS. (as of summer 2003, the TFS had only been partially released by NCES and data presented here from that cycle are preliminary estimates). DEMAND FOR TEACHERS HAS RISEN What do the data tell us about school staffing problems and teacher shortages? The data show that the conventional wisdom on teacher shortages is correct in some respects. Consistent with shortage predictions, data from SASS and other NCES data sources show that demand for teachers has indeed increased in recent years. Since 1984, student enrollments have increased, most schools have had job openings for teachers, and the size of the teaching workforce (K-12) has increased, although the rate of these increases began to decline slightly in the late 1990s (Gerald & Hussar, 1998; Snyder & Hoffman 2001, pp. 11). Most importantly, many schools with teaching openings have experienced difficulties with recruitment. Overall, the data show that in the school year, 58% of all schools reported at least some difficulty filling one or more teaching job openings, in one or more fields. However, the data also show that in any given field less than half of the total population of schools actually experienced recruitment problems (see Figure 1). For instance in , 54% of secondary schools had job openings for English teachers and about one half of these indicated they had at least some difficulty filling these openings representing one quarter of all secondary schools. Similarly, 54% of secondary schools had job openings for math teachers and about four fifths of these indicated they had at least some difficulty filling these math openings representing about 40% of all secondary schools. Likewise, 45% of secondary schools had job openings for special education 6

9 teachers and about three quarters of these indicated they had at least some difficulty filling these openings representing 34% of secondary schools. 1 Figure 1. Percent Secondary Schools with Vacancies and Percent with Difficulties Filling those Vacancies, English Mathematics Special Education Life Science Physical Science Social Studies Music/Art ESL/ESOL/Bilingual Foreign Language Percent Vacancies Hiring Difficulties HOW ADEQUATE IS THE SUPPLY OF TEACHERS? While demand has increased and many schools have had hiring difficulties, the data do not show, contrary to the conventional wisdom, that there is overall an insufficient supply of teachers being produced. National data on the overall supply of teachers trained, licensed, and certified each year are difficult to obtain. One source is NCES Integrated Postsecondary Educational Data System (IPEDS). This source collects data on the numbers of post-secondary degree completions by field and by year. The IPEDS data indicate that, for example, at the end of the academic year there were over 220,000 new recipients of education degrees at the undergraduate and graduate levels (IPEDS 2001). But, the SASS data show that only about 86,000 of those hired for the following school year were drawn from this group of recent college graduates. Indeed, large proportions of those who train to become teachers do not ever become teachers. For example, data from the nationally representative 1993 Baccalaureate and Beyond Survey show that, of new recipients of bachelor s of education degrees who graduated in 1993, after one year out of college only 42% had taught and after four years out of college only 58% had taught. Many of those 7

10 who decided not to pursue teaching indicated that they needed more education, or wanted another occupation (for a more detailed presentation of the B&B data, see Henke et al. 2000). In short, the data appear to indicate that, overall, there are more than enough prospective teachers produced each year in the U.S. But, there are also some important limitations to these data. An overall surplus of newly trained teachers does not, of course, mean there are sufficient numbers of graduates produced in each field. A large proportion of education degree completions are in elementary education. The data are unclear on whether a sufficient quantity of teachers are produced each year in such fields as math, science, and special education. But, on the other hand, the IPEDS database on degree completions underestimates the supply of newly qualified teachers because it does not include recipients of non-education undergraduate degrees who also completed the requirements for certification. Moreover, newly qualified candidates, as counted in the IPEDS data, are only one source of new hires in schools. Far more of those newly hired into schools each year are from what is often referred to as the reserve pool. These include delayed entrants those who completed teacher training in prior years, but who have never taught, and re-entrants former teachers who left teaching for a period to later return. The addition of these other types and sources of teachers could well mean that there are more than enough teachers produced each year. However, from an organizational perspective, the key question is not whether the overall national supply of teachers is adequate or inadequate, instead it is which schools have staffing problems and teacher supply and demand imbalances? Even in the same jurisdiction, the degree of staffing problems varies greatly among different types of schools, and sites ostensibly drawing from the same teacher supply pool can have significantly different staffing scenarios. Some analysts have found, for example, that in the same metropolitan area in the same year some schools have extensive waiting lists of qualified candidates for their teaching job openings, while other nearby schools have great difficulty filling their teaching job openings with qualified candidates (National Commission on Teaching, 1997). Consistent with this, I have found in an analysis of variance of the SASS data that the variation in school hiring difficulties is far greater within, than between, states. 2 This suggests that to be fully understood imbalances between demand and supply must be examined at the level of the organization an issue to which I return. Figure 2. Percent Annual Teacher Turnover Percent Movers Leavers 8

11 THE IMPORTANCE OF TEACHER TURNOVER FOR SCHOOL STAFFING PROBLEMS There is another problem with the conventional wisdom on shortages. The data show that the demand for new teachers and subsequent staffing difficulties are not primarily due to student enrollment and teacher retirement increases, as widely believed, but these are largely due to teacher turnover teachers moving from or leaving their teaching jobs and most of this turnover has little to do with a graying workforce. The TFS data show that teaching has a relatively stable annual turnover rate: 14.5% in ; 13.2% in ; 14.3% in ; 15.7% in (See Figure 2). There are two types of total turnover included in Figure 2: movers those who move to teaching jobs in other schools (often and hereafter referred to as teacher migration) and leavers those who leave the teaching occupation altogether (often and hereafter referred to as teacher attrition). Total teacher departures are fairly evenly split between them. Much of the existing research on teacher turnover does not include the former. Teacher cross-school migration is a form of turnover that does not decrease the overall supply of teachers because departures are simultaneously new hires. As a result, many assume that teacher migration does not contribute to the problem of staffing schools and to overall shortages. From a macro and system level of analysis, this is probably correct and for this reason educational researchers have often de-emphasized or excluded movers. However, from an organizational perspective and from the viewpoint of those managing at the school-level, movers and leavers have the same effect in either case it results in a decrease in staff, which usually must be replaced. Hence, research on employee turnover in other occupations and organizations almost always includes both movers and leavers and for this reason I include them here. As illustrated in Figure 2, adopting a system-level or an organizational-level of analysis makes a difference if one excludes cross-school moves, total turnover would appear far less than it is from the viewpoint of those managing schools. One question that naturally arises concerns how rates of teacher turnover from schools compare to those in other organizations, occupations, and industries. There has never been much empirical research on this issue. In earlier work I compared the TFS data with data from one of the better known sources of information on employee turnover in a range of occupations and organizations, the Bureau of National Affairs (2002). This comparison showed that teaching has a slightly higher annual turnover rate than the nationwide level for total employee departures which over the past decade averaged 11.9% annually (Ingersoll 2001a). But, this comparison provided only a crude benchmark. The BNA data represent an organizational rate of employee turnover, which ostensibly includes the whole range of employees on the payroll from clerical staff to senior management within a given organization. Far more informative would be comparisons of occupational rates of turnover, i.e., how do rates of teacher turnover compare to those in other occupations? But, such data are difficult to obtain and suffer from serious issues of measure comparability (Price 1977, 1997). 3 In more recent work I have found that, as one might expect, teaching has higher turnover than some higher-status professions (professors at 9.3%; technology and scientific professionals from 3.6% to 9.2%), about the same as other female semi-professions (nurses at 18%) and less turnover than some lower-status, lower-skill occupations (federal clerical workers at 30%). 4 But, from an organizational perspective the key question is not whether teaching has higher or lower turnover than other occupations, but rather is teacher turnover a problem for schools themselves. The data indicate it is. There is a strong link between teacher turnover and the difficulties schools have adequately staffing classrooms with 9

12 qualified teachers as shown in Table 1 which presents data on the flows in and out of schools from all four cycles of the SASS/TFS data. Reading down the column for the school year, for example, Table 1 indicates there were just under three million teachers in the K-12 education system that year, including both public and private schools. About 377,135 of these teachers entered their schools at the beginning of the school year. Of these, 192,550 had not taught the prior year. This latter group included newly qualified candidates fresh out of college, delayed entrants who had completed their training in a prior year but had not previously taught, and re-entrants who had taught previously, stopped for a while and then returned. Another 184,585 of these hires to schools had moved from another school. By the following school year, 417,588 teachers had moved from or left their school jobs. Just under half of these departures 204,680 moved to other schools to teach. Another 212,908 left the occupation altogether. Of the latter, 50,242 were retirees. Table 1 documents two important points. First, the data show that the demand for new teachers is not primarily due to student enrollment increases, nor to teacher retirement increases, but to pre-retirement teacher turnover. That is, most of the hiring of new teachers is simply to fill spots vacated by teachers who just departed. For instance, about 191,000 individuals entered teaching at the beginning of the school year. However, by the following school year 12 months later, about 174,000 teachers (equivalent to 91% of those just hired) left the occupation altogether. At the beginning of the , three years later, about 192,500 teachers entered teaching, but by the following school year, about 213,000 (equivalent to 110% of those just hired) left the occupation. At the beginning of the , six years later, about 232,000 teachers entered teaching, but by the following school year, about 287,000 (equivalent to 124% of those just hired) left the occupation. Table 1: Trends in Teacher Flows In and Out of Schools School Year School Year School Year School Year 1.) Total Teaching Force during school year 2,630,335 2,915,774 2,939,659 3,451,316 2.) Total Hires - at beginning of school year 361, , , , 861 A.) Entrants 178, , , ,232 B.) Movers from other schools 183, , , ,629 3.) Total Departures by following school year 390, , , ,778 Notes: A.) Movers to other schools 218, , , ,408 B.) Leavers from occupation 172, , , ,370 Retirees 35,179 47,178 50,242 NA Entrants: includes new, delayed, and re-entrants. This refers to those who did not teach the prior year; some did teach in the past. These data are calculated at the level of the school. Hence, hires and departures refer to those newly entering or departing a particular school. Movers includes transfers among schools within districts. Reassignments within a school are not defined as hires or as departures. As of summer 2003, teacher retirement data for the 2000 school year were not yet available. 10

13 Second, although teacher retirements have increased in recent years, they account for only a small portion of the above total turnover. For example, from 1994 to 1995 there were about 50,000 retirees, accounting for only 24% of the 213,000 leavers and only 12% of the total turnover of 417,588 during that period. In sum, the data show that the demand for new teachers, and subsequent staffing difficulties, is not primarily due to student enrollment increases, nor to teacher retirement increases, but to preretirement teacher turnover. That is, most of the hiring of new teachers is simply to fill spots vacated by teachers who just departed. And, most of those departing are not doing so because of gray hair. THE REVOLVING DOOR It is also important to note that teaching is a relatively large occupation it represents 4% of the entire civilian workforce. There are, for example, over twice as many K-12 teachers as registered nurses and five times as many teachers as either lawyers or professors (U.S. Bureau of the Census, 2002). 5 The sheer size of the teaching force combined with its relatively high annual turnover means that there are large flows in, through, and out of schools each year. The image that these data suggest is one of a revolving door, which I have tried to capture in Figure 3. It shows that for the school year, 534,861 teachers entered schools, while by the following school year an even larger number 539,778 had moved from or left their schools. 6 Hence, in a 12-month period over one million teachers almost a third of this relatively large workforce were in job transition into, between, or out of schools. This revolving door is a major factor behind school staffing problems. Table 1 and Figure 3 also provide a context to interpret the widely used statistic, introduced earlier, that the nation will need to hire at least two million teachers over the next 10 years. This statistic was drawn from an NCES analysis (Gerald & Hussar 1998; Hussar 1998) that projected the numbers of teachers that would need to be hired from 1998 to 2008 in order to replace those who had left teaching and to account for student enrollment increases. These analyses themselves did not examine supply, nor changes in hiring needs over time. But, a wide variety of commentators, researchers, and policy makers have interpreted this statistic to mean that hiring two million new teachers is an unusually large number and assumed to be evidence that we face an alarmingly inadequate supply of new teachers being produced. A close look at the data reveals that neither is the case. THE IMPORTANCE OF TEACHER TURNOVER FOR ORGANIZATIONS Of course, not all teacher turnover is detrimental. There is an extensive research literature on employee turnover conducted by those who study organizations and occupations in general (e.g., Price 1977, 1989; Mueller & Price, 1990; Bluedorn, 1982; Halaby & Weakliem, 1989; Hom & Griffeth, 1995; Kalleberg & Mastekaasa, 1998; March & Simon, 1958; Mobley, 1982; Steers & Momday, 1981). On the one hand, researchers in this tradition have long held that a low level of employee turnover is normal and efficacious in a well-managed organization. Too little turnover of employees is tied to stagnancy in organizations; effective organizations usually both promote and benefit from a limited degree of turnover by eliminating low-caliber performers and bringing in new blood to facilitate innovation. Moreover, some job and career changes are, of course, normal and inevitable in any occupation. On the other hand, researchers in this tradition have also long held that high levels of employee turnover are both cause and effect of performance problems in organizations. 11

14 Figure 3. Numbers of Teachers in Transition Before, During, and After the School Year. Not in Transition 2,376,677 Entrants at Beginning 534,861 Departures at End 539,778 Organizational analysts have also noted that the consequences of employee turnover vary among different types of employees and among different types of organizations. Labor process analysts, for instance, have argued that a major issue, from the viewpoint of organizational management, is the extent to which the organization is, or is not, dependent on particular types of employees and, hence, vulnerable to the disruption caused by their turnover (e.g. Braverman 1974; Burawoy 1979; Edwards 1979). For just this reason the issue of employee substitutability, or the ease with which organizations can replace employees, is a central concern in organizational management and a central theme in organizational research. In this perspective, employee turnover is especially consequential for work that involves uncertain and non-routine technologies and which requires extensive interaction among participants. Such organizations are often unusually dependent upon the commitment and cohesion of employees and, hence, especially vulnerable to turnover (e.g., Burns & Stalker, 1961; Kanter, 1977; Likert, 1967; Porter, Lawler & Hackman, 1975; Turner & Lawrence, 1964; Walton, 1980). Schools are an example of this type of organization. Education theory and research have long shown that, while education is a mass industry involving large complex formal organizations, in important ways schools do not fit standard input-output, economic-production models in either theory or practice (Bidwell, 1965; Lortie, 1975; Ingersoll, 2003a). The raw materials in schools are children and youth; the technology of teaching and learning is often uncertain, ambiguous, and non-routine; and the product is youngsters growth. As a result while schools in some ways resemble economic-production organizations, in other ways they resemble another kind of institution altogether the family. Student test outcomes are one of the important output functions of school production. But not surprisingly, similar to families the presence of a positive sense of community, belongingness, communication, and cohesion among members has long been held by education theory and research to be one of the most important indicators and aspects of effective schools (e.g., Durkheim, 1925/1961; Waller, 1932; Parsons 1959; Grant, 1988; Coleman & Hoffer, 1987; Kirst 1989; Rosenholtz 1989). Hence, from an organizational perspective, some teacher turnover, especially of ineffective teachers, is necessary and beneficial. But from this perspective, turnover 12

15 of teachers from schools is of concern not simply because it may be an indicator of sites of potential staffing problems and so-called teacher shortages, but because of its relationship to school cohesion and, in turn, performance. Moreover, from this perspective this relationship runs both directions. That is, high rates of teacher turnover are of concern not only because they may be an outcome indicating underlying problems in how well schools function, but also because they can be disruptive, in and of themselves, for the quality of school community and performance. Some of these costs and consequences of turnover are more easily measured and quantified than others. In contrast to the corporate sector, however, there has been very little attention paid to the costs and consequences of employee turnover in education. One notable exception was a recent attempt to quantify the costs of teacher turnover in Texas this study concluded these costs run into the hundreds of millions of dollars each year to the state. (Texas Center for Educational Research, 2000). TEACHER AND SCHOOL DIFFERENCES IN TURNOVER The data also show that the revolving door varies greatly among different kinds of teachers and different kinds of schools. As found in previous research (Murnane et al. 1991; Huling-Austin 1990; Hafner & Owings 1991), the SASS data show that teaching is an occupation that loses many of its newly trained members very early in their careers long before the retirement years. I used these data to provide a rough estimate of the cumulative attrition of beginning teachers from the occupation in their first several years of teaching. The data suggest that after just five years, between 40 and 50% of all beginning teachers have left teaching altogether. (see Figure 4). 7 Of course, not all of this attrition results in a permanent loss of teachers. One form of this revolving door is represented by temporary attrition teachers who leave teaching but return in later years, as discussed earlier (also see Murnane et al. 1991). But again, from the viewpoint of those managing at the school-level, temporary and permanent attrition have the same effect in either case it results in an immediate decrease in staff, which usually must be replaced. Annual teacher turnover also varies according to the teachers main field. Math, science, and elementary special education teachers have higher rates of turnover, and social studies and English have lower rates (see Figure 5). Moreover, a number of studies have found that teachers with higher ability, as measured by test scores such as the SAT, the National Teacher Exam, and teacher licensure tests, are more likely to turn over (e.g., Weaver 1983; Murnane et al. 1991; Schlecty & Vance 1981; Stinebrickner 2001; Henke et al. 2000). The data also show that the revolving door varies greatly among different kinds of schools, as illustrated in Figure 6. 8 For example, high-poverty public schools have far higher turnover rates than do more affluent public schools. Urban public schools have slightly more turnover than do suburban and rural public schools. Private schools have higher turnover rates than public schools, but there are also large differences among private schools. On one end of the continuum lie larger private schools with among the lowest average turnover rate about 13.5%. On the other end of the continuum lie smaller private schools with among the highest average levels about 22%. 13

16 Figure 4. Beginning teacher attrition (cumulative percent teachers having left teaching occupation, by years of experience). After 1 yr. 14 After 2 yrs. 24 After 3 yrs. 33 After 4 yrs. 40 After 5 yrs Percent Figure 5. Percent annual teacher turnover, by field ( ). All Teachers Mathematics Science Special Educ. Elementary Elementary Special Educ. Secondary Voch. Tech. English Social Studies Percent Movers Leavers 14

17 Figure 6. Percent annual teacher turnover, by selected school characteristics ( ). All Teachers Public School Teachers Low Poverty High Poverty Rural Suburban Urban Large Small Private School Teachers Catholic Other Religious Non-sectarian Large Small Percent Movers Leavers THE SOURCES OF TEACHER TURNOVER These data raise another important set of questions: why do teachers depart at relatively high rates and why are these rates so dramatically different among schools? To answer these questions I conducted multivariate statistical analyses of data from different cycles of SASS/TFS to determine which characteristics of teachers and schools are associated with the likelihood of teacher turnover, after controlling for background factors (Ingersoll 2001b). I also examined data on the reasons teachers themselves give for their turnover. Such self-report data are useful because those departing are, of course, often in the best position to know the reasons for their turnover. But, such self-report data are also retrospective attributions, subject to bias and, hence, warrant caution in interpretation. Nevertheless, I found a great deal of consistency among these different types of data and from different cycles of the survey. The following section summarizes my principal findings. Along with the annual rates of turnover, Table 2 presents self-report data on teachers reasons for both migration and attrition. In addition, for all teachers who departed because of job dissatisfaction, the bottom portion of the table presents data on the reasons for their dissatisfaction. 9 (definitions of these reasons can be found in the Appendix) 15

18 Table 2. Percent Teacher Turnover and Percent Teachers Giving Various Reasons for their Turnover, All Movers Leavers Rates of Turnover Reasons for Turnover Retirement School Staffing Action Family or Personal To Pursue other Job Dissatisfaction Reasons for Dissatisfaction Poor Salary Poor Administrative Support Student Discipline Problems Lack of Faculty Influence & Autonomy Poor Student Motivation No Opportunity for Professional Advancement Inadequate Time to Prepare Intrusions on Teaching Time Class Sizes too Large Contrary to conventional wisdom, retirement is not an especially prominent factor. The latter was listed by only 25% of leavers and 13% of total departures. School staffing cutbacks due to lay-offs, terminations, school closings, and reorganizations account for a larger proportion of turnover than does retirement. These staffing actions more often result in migration to other teaching jobs rather than leaving the teaching occupation altogether. But, the data also show that, overall, staffing actions, like retirement, account for only a small portion of total turnover from schools about 20%. A third category of turnover that for personal reasons includes departures for pregnancy, child rearing, health problems, and family moves. These account for more turnover than either retirement or staffing actions and they are probably common to all occupations and all types of organizations. The two final sets of reasons are directly related to the organizational conditions of teaching. Individually each of these categories accounts for more turnover than does retirement. Together these are the most prominent source of turnover. Almost half of all departures report as a reason either job dissatisfaction or the desire to pursue a better job, another career, or to improve career opportunities in or out of education. Of those who depart because of job dissatisfaction, most often link their turnover to low salaries, lack of support from the school administration, student discipline problems, poor student motivation, and lack of teacher influence over decision-making. In general, similar kinds of dissatisfactions lie behind both teacher migration and teacher attrition. Interestingly, several factors stand out as not major reasons behind turnover, according to those who departed: large class sizes, intrusions on classroom time, and lack of planning time. In sum, the data indicate that teachers depart their jobs for a variety of reasons. Retirement accounts for a relatively small number of total departures, a moderate number of departures are due to school staffing actions, a large proportion indicate 16

19 they depart for personal reasons, and a large proportion also report they depart either because they are dissatisfied with their jobs or in order to seek better jobs or other career opportunities. I have found these reasons for turnover to be highly consistent across different types and cycles of the data, across different kinds of schools, and across different subsets of teacher turnover. These findings are important because of their policy implications. Unlike explanations that focus on external demographic trends out there, these findings suggest there is a role in here for the internal organization and management of schools. IMPLICATIONS FOR POLICY It is widely believed that shortfalls of teachers resulting primarily from two converging demographic trends increasing student enrollments and increasing teacher retirements are leading to problems staffing schools with qualified teachers and will, in turn, lower educational performance. In response school districts, states, and the federal government have developed a variety of recruitment initiatives designed to recruit more candidates into teaching. However worthwhile these efforts may be, the data suggest that, alone, they will not solve school staffing problems. The data suggest that school staffing problems are not solely or even primarily due to teacher shortfalls resulting from either increases in student enrollment or increases in teacher retirement. In contrast, the data suggest that school staffing problems are to a large extent a result of a revolving door where large numbers of teachers depart teaching for reasons other than retirement. From the framework of supply and demand theory, the data show that the problem is not primarily shortages, in the sense of an insufficient supply of teachers being recruited and trained. Some economists would still call this a shortage in the technical sense that there is an inadequate quantity of teachers supplied those willing to continue to offer their services at a given wage and given working conditions in given schools. This diagnostic and terminological distinction has crucial implications for prescription. It is also a distinction that is almost always overlooked in research and policy on the teacher shortage. Supply and demand theory holds that where the quantity of teachers demanded is greater than the quantity of teachers supplied, there are two basic policy remedies: increase the quantity supplied or decrease the quantity demanded. The first approach the traditionally dominant approach is to increase the quantity of teachers supplied through recruitment. However, this analysis cautions that recruitment programs alone will not solve the staffing problems of schools if they do not also decrease turnover. States such as California, where class-size-reductions have strained the supply of new teachers pose exceptions. But, for just these reasons, California, like other states, must pay close attention to retention. In short, recruiting more teachers will not solve the teacher crisis if 40 to 50% of such teachers then leave within five years. The image that comes to mind is a bucket rapidly losing water because of holes in the bottom. Pouring more water into the bucket will not be the answer if the holes are not first patched. Recruitment and other supply-side solutions may not only fail to solve the problem but could also make the situation worse. If recruitment strategies involve lowering teacher standards, or if the effect of increasing teacher supply is to deflate salaries or erode working conditions, then these measures may simply exacerbate the root factors behind school staffing problems. This situation is analogous to aspects of management-labor conflict in industry. Critics of business practice argue that industrialists have long used labor supply recruitment as a strategy to undermine worker and union efforts to improve 17

20 working conditions and wages (e.g. Braverman 1974; Burawoy 1979; Edwards 1979). For example, by bringing in immigrant laborers from eastern and southern Europe at the turn of the 19th century, industrialists, the critics hold, were able to keep wages down, undermine union power, and increase profits. One of the downsides with this strategy, from a management perspective, is that it can decrease employee quality and increase employee turnover. Hence, one of the objectives behind the design of the assembly-line model of production used in industry was to increase the ease of substitutability and, hence, insulate the organization from disruption caused by employee turnover. Similarly, social scientists have long characterized K-12 teaching as a lower status, easy-in/easy-out, high turnover occupation that has relied historically on recruitment, and not retention, to solve its staffing problems (e.g., Tyack 1974; Lortie 1975). Since the inception of the public school system in the late 19th century, teaching was socially defined and treated as a temporary line of work suitable for women, prior to their real career of child rearing. For men, teaching was socially defined as a stepping stone, prior to their real career in one of the male-dominated skilled blue-collar occupations or white-collar professions. Indeed, historically there was an ambivalence toward persistors in teaching, especially males who had to account for why they continued to be merely a teacher. Low pre-service training standards and requirements, relatively unselective entry criteria, and front-loaded salaries that paid newcomers relatively high salaries compared to veterans all tended to favor recruitment over retention. Moreover, low pay, isolated job conditions, little professional autonomy, and a faint sense of a career ladder all undermined longerterm commitment to teaching as a career and profession. Attempts to upgrade the status of the occupation through more rigorous training and licensing standards or more selective entry gates often resulted in decreases in male entrants to teaching, who were more attracted to occupations with better rewards attached to rigorous standards (Strober & Tyack 1980). It appears that school districts continue to favor teacher recruitment strategies for many of the same reasons and with many of the same consequences. By widening the entry gate and increasing the quantity of teachers supplied, districts are able to control labor costs and, hence, control local property taxes. The downside of this strategy in schools, as in industry, is that it can decrease employee quality and increase employee turnover. Treating workers as interchangeable, expendable, low-skill workers cuts some expenses, but it is not cost-free. If turnover is at the root of school staffing problems and if the quality of the teaching job is a large factor behind turnover, then policies that further erode the low status of teaching, that undermine salary increases, or that undermine working conditions may simply backfire by increasing turnover. In short, the data suggest that school staffing problems are rooted in the way schools are organized and the way the teaching occupation is treated and that lasting improvements in the quality and quantity of the teaching workforce will require improvements in the quality of the teaching job. 18